| stevebaker | lifeless: can you cast your mind back to your preserve_ephemeral heat patch? | 00:00 |
|---|---|---|
| *** slong has joined #heat | 00:03 | |
| *** gokrokve has quit IRC | 00:04 | |
| *** slong_ has quit IRC | 00:04 | |
| *** gokrokve has joined #heat | 00:04 | |
| * SpamapS would help but EOD | 00:05 | |
| SpamapS | stevebaker: we should probably adapt to pass with/without that arg present in novaclient, since it is optional. | 00:06 |
| SpamapS | anyway.. bad timing.. have to run | 00:06 |
| *** gokrokve has quit IRC | 00:09 | |
| lifeless | stevebaker: yes | 00:13 |
| lifeless | stevebaker: wassup ? | 00:13 |
| *** mkollaro1 has quit IRC | 00:23 | |
| stevebaker | lifeless: did you actually have issues passing preserve_ephemeral=False to older novaclient? or was it defensive? | 00:28 |
| *** rongze has joined #heat | 00:35 | |
| lifeless | stevebaker: its a new keyword parameter | 00:36 |
| lifeless | stevebaker: mmm, but yes the older does have **kwargs there | 00:37 |
| lifeless | stevebaker: however, it gets passed down to _action as a **kwargs too | 00:38 |
| lifeless | stevebaker: it's possibly harmless to pass it to earlier versions. Why? | 00:38 |
| *** topol has joined #heat | 00:39 | |
| *** rongze has quit IRC | 00:40 | |
| stevebaker | lifeless: because git master novaclient changes the method signature that needs to be mocked, and I can't see a way of getting mox to handle old and new novaclient https://bugs.launchpad.net/heat/+bug/1271367 | 00:41 |
| stevebaker | lifeless: one easy fix is to always pass preserve_ephemeral | 00:41 |
| *** giulivo has quit IRC | 00:42 | |
| lifeless | huh | 00:43 |
| lifeless | I don't see the change in master | 00:43 |
| lifeless | how did that land without breaking heat in the gate? | 00:45 |
| *** kebray_ has quit IRC | 00:45 | |
| stevebaker | do we test against released novaclient? That would explain it | 00:45 |
| *** sgran has quit IRC | 00:45 | |
| lifeless | or your unit tests are not run as part of landing novaclient changes | 00:45 |
| lifeless | but since you run novaclient code, they should be | 00:46 |
| lifeless | want a bug about that? | 00:46 |
| lifeless | anyhow | 00:46 |
| lifeless | are heat merge proposals now failing check ? | 00:46 |
| stevebaker | not yet. just locally for me with git novaclient | 00:47 |
| stevebaker | it hadn't occured to me that we should gate heat unit tests on client changes | 00:47 |
| lifeless | the rule is 'if your tests can be broken by a change to X, you should tell X that before X can land their change' | 00:47 |
| lifeless | in general | 00:48 |
| lifeless | so either isolate heat more | 00:48 |
| lifeless | or mutually gate | 00:48 |
| lifeless | anyhow | 00:48 |
| lifeless | yes you may be testing with pip novaclient | 00:48 |
| lifeless | which would mean that when novaclient releases you'll have a firedrill | 00:48 |
| *** sgran has joined #heat | 00:48 | |
| stevebaker | yep. | 00:50 |
| lifeless | stevebaker: ok, so - sure, passing preserve_ephemeral unconditionally seems like a reasonable workaround | 00:51 |
| lifeless | its a it uck aesthetically but whatever :) | 00:51 |
| stevebaker | hey, it makes *our* code cleaner ;) | 00:51 |
| * stevebaker kicks down | 00:51 | |
| *** IlyaE has joined #heat | 00:53 | |
| *** topol has quit IRC | 00:56 | |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Always specify preserve_ephemeral on server rebuild https://review.openstack.org/68294 | 01:06 |
| *** blamar has joined #heat | 01:08 | |
| *** pvaneck has quit IRC | 01:11 | |
| *** blomquisg has quit IRC | 01:17 | |
| *** mestery has quit IRC | 01:18 | |
| *** rpothier has joined #heat | 01:29 | |
| *** blomquisg has joined #heat | 01:30 | |
| *** Linz has quit IRC | 01:31 | |
| *** Linz has joined #heat | 01:31 | |
| *** harlowja has joined #heat | 01:33 | |
| *** nosnos has joined #heat | 01:35 | |
| *** pablosan_ has joined #heat | 01:43 | |
| *** rongze has joined #heat | 01:46 | |
| *** Linz has quit IRC | 01:46 | |
| *** pablosan has quit IRC | 01:46 | |
| *** kebray_ has joined #heat | 01:46 | |
| *** Linz has joined #heat | 01:51 | |
| *** Linz has quit IRC | 01:51 | |
| *** Linz has joined #heat | 01:51 | |
| *** IlyaE has quit IRC | 02:01 | |
| *** renlt has joined #heat | 02:05 | |
| *** rpothier has quit IRC | 02:08 | |
| *** rpothier has joined #heat | 02:09 | |
| *** dims has quit IRC | 02:11 | |
| *** kebray_ has quit IRC | 02:19 | |
| *** michchap has quit IRC | 02:24 | |
| *** michchap has joined #heat | 02:25 | |
| *** andrew_plunk has joined #heat | 02:28 | |
| radix | zaneb: thanks for the review | 02:36 |
| zaneb | radix: does that explain it better? | 02:37 |
| radix | zaneb: yep :) | 02:37 |
| zaneb | cool :) | 02:37 |
| radix | zaneb: by the way, since you're here, I can tell you about the change to the changing_instances function | 02:37 |
| radix | zaneb: basically, it relied on comparing the str() of dicts, and that's broken | 02:37 |
| radix | and unfortunately, there is no "partition" function for lists in Python, so I had to write some for loops with conditionals | 02:37 |
| zaneb | broken in what sense? | 02:38 |
| radix | str(dict) has an unreliable order of keys. I ran into it in testing | 02:38 |
| radix | that shourd probably be another patch :) | 02:38 |
| radix | should* | 02:39 |
| zaneb | hmmm | 02:39 |
| zaneb | that's surprising, because strings should have the same hashes | 02:40 |
| radix | it's not the hashing that's the problem, it's the str(dict) | 02:40 |
| radix | it will return keys in different orders | 02:40 |
| radix | because dict keys aren't ordered | 02:41 |
| zaneb | well, they're ordered in the order they're found in the hash table ;) | 02:41 |
| radix | right, which is not the sort order of the keys, or in any way reliable | 02:41 |
| radix | I actually had test failures because of it | 02:42 |
| zaneb | but I guess that order depends on the size of the hash table? | 02:42 |
| zaneb | intersting | 02:42 |
| radix | and maybe the circumstances of the addition of keys to the dict? | 02:42 |
| zaneb | yes, I think depending on how you added keys to the dict that could affect the size of the hash table it's using | 02:43 |
| zaneb | and that in turn could affect the key order | 02:43 |
| zaneb | tricky | 02:43 |
| radix | another way to fix that would be to implement an immutable mapping class with a value-based __hash__ | 02:43 |
| radix | and still use the same set logic | 02:43 |
| radix | anyway, I can file a separate bug for taht | 02:44 |
| zaneb | radix: I suggest using json.dump(template, sort_keys=True) | 02:44 |
| zaneb | but yes, separate bug++ :) | 02:44 |
| radix | oh, that's probably good too :) | 02:45 |
| zaneb | there was a frozendict (immutable dict) class proposed for the standard library once | 02:46 |
| zaneb | BDFL rejected it though :/ | 02:46 |
| radix | yeah :( | 02:46 |
| radix | the rationale was pretty... eh, I disagreed with it, I'll say that :) | 02:46 |
| zaneb | I haven't actually read it | 02:46 |
| zaneb | anyway, not relying on Python serialisation is probably a Good Thing | 02:47 |
| zaneb | at some point I want to create a ResourceTemplate class | 02:48 |
| radix | please don't make it subclass dict :P | 02:48 |
| zaneb | that would just be a template snippet for a resource | 02:48 |
| *** ktbenton has joined #heat | 02:48 | |
| *** ktbenton has quit IRC | 02:48 | |
| zaneb | right, yeah the point would be to *not* subclass dict | 02:48 |
| radix | :) | 02:48 |
| zaneb | we need some sort of abstraction for HOT vs cfn | 02:49 |
| zaneb | there is too much poking around in dicts by the resources | 02:49 |
| zaneb | anyway, it could be a proper hashable | 02:50 |
| *** erkules_ has joined #heat | 02:51 | |
| stevebaker | zaneb: do we have a set of blueprints for typing of template data? | 02:51 |
| zaneb | stevebaker: I'm gonna say yes | 02:52 |
| zaneb | by 'typing', you mean... | 02:52 |
| zaneb | separating cfn & HOT? | 02:52 |
| stevebaker | so that our code isn't full of rumaging through dicts | 02:52 |
| *** erkules has quit IRC | 02:53 | |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: HOT templates get_attr allows extra attributes https://review.openstack.org/67562 | 02:55 |
| *** andrew_plunk has quit IRC | 03:00 | |
| *** slong_ has joined #heat | 03:01 | |
| *** slong has quit IRC | 03:01 | |
| zaneb | stevebaker: there is this one https://blueprints.launchpad.net/heat/+spec/schema-code-consolidation | 03:09 |
| zaneb | stevebaker: but you make a good point, I'm not sure if there are bps for the rest of the stuff that has to happen | 03:09 |
| stevebaker | something we just need to chip away at I guess | 03:10 |
| *** liang has joined #heat | 03:11 | |
| openstackgerrit | Lee Li proposed a change to openstack/python-heatclient: Reuse Resource from oslo https://review.openstack.org/67120 | 03:12 |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Native Pseudo Parameters https://review.openstack.org/66640 | 03:13 |
| *** ramishra has joined #heat | 03:13 | |
| *** erkules_ is now known as erkules | 03:18 | |
| openstackgerrit | Vijendar Komalla proposed a change to openstack/python-heatclient: python-heatclient stack-abandon support https://review.openstack.org/62925 | 03:27 |
| openstackgerrit | Vijendar Komalla proposed a change to openstack/heat: Implement adopt-stack https://review.openstack.org/62730 | 03:28 |
| openstackgerrit | Vijendar Komalla proposed a change to openstack/heat: Implement adopt-stack for nested stacks https://review.openstack.org/64720 | 03:28 |
| openstackgerrit | Christopher Armstrong proposed a change to openstack/heat: Make LB-updating in rolling update more reliable https://review.openstack.org/68311 | 03:32 |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Prevent access Parameters key in template dict https://review.openstack.org/68312 | 03:38 |
| *** slong has joined #heat | 03:43 | |
| *** slong_ has quit IRC | 03:43 | |
| *** harlowja is now known as harlowja_away | 03:54 | |
| *** gokrokve has joined #heat | 03:55 | |
| openstackgerrit | Lee Li proposed a change to openstack/python-heatclient: Remove unused method 'bool_from_string' from utils https://review.openstack.org/68317 | 03:58 |
| openstackgerrit | Liang Chen proposed a change to openstack/heat: Marconi message queue resource implementation https://review.openstack.org/61036 | 04:23 |
| *** akuznetsov has joined #heat | 04:25 | |
| *** spzala has quit IRC | 04:28 | |
| *** gokrokve has quit IRC | 04:28 | |
| liang | therve, there? | 04:28 |
| *** gokrokve has joined #heat | 04:29 | |
| *** arbylee has joined #heat | 04:45 | |
| *** nkhare has joined #heat | 04:48 | |
| *** arbylee has quit IRC | 04:49 | |
| *** rongze has quit IRC | 04:52 | |
| *** gokrokve has quit IRC | 04:53 | |
| *** faramir1 has joined #heat | 04:57 | |
| *** faramir1 has quit IRC | 05:03 | |
| *** chandankumar_ has joined #heat | 05:08 | |
| *** rongze has joined #heat | 05:23 | |
| *** gokrokve has joined #heat | 05:25 | |
| openstackgerrit | Angus Salkeld proposed a change to openstack/heat: Update olso gettextutils https://review.openstack.org/67758 | 05:26 |
| openstackgerrit | Angus Salkeld proposed a change to openstack/heat: Update oslo db https://review.openstack.org/67759 | 05:26 |
| openstackgerrit | Angus Salkeld proposed a change to openstack/heat: Update oslo modules https://review.openstack.org/67757 | 05:26 |
| *** rongze has quit IRC | 05:28 | |
| *** rongze has joined #heat | 05:35 | |
| *** rongze has quit IRC | 05:40 | |
| openstackgerrit | A change was merged to openstack/heat: new method 'HeatTestCase.patchobject' https://review.openstack.org/67234 | 05:45 |
| openstackgerrit | A change was merged to openstack/heat: heat_keystoneclient migrate auth_token/service_catalog to v3 API https://review.openstack.org/66649 | 05:51 |
| openstackgerrit | A change was merged to openstack/heat: Turn block_device_mapping.volume_size to int https://review.openstack.org/66804 | 05:51 |
| *** e0ne has joined #heat | 05:53 | |
| openstackgerrit | A change was merged to openstack/heat: Fix incorrect resource types in docs/comment/code https://review.openstack.org/67404 | 05:53 |
| openstackgerrit | A change was merged to openstack/heat: Refactor MySQL long text support https://review.openstack.org/67406 | 05:54 |
| openstackgerrit | A change was merged to openstack/python-heatclient: Raise traceback on error when using CLI and -debug https://review.openstack.org/66136 | 05:54 |
| *** e0ne has quit IRC | 05:54 | |
| *** IlyaE has joined #heat | 06:02 | |
| *** blamar has quit IRC | 06:08 | |
| *** IlyaE has quit IRC | 06:10 | |
| *** aignatov_ is now known as aignatov | 06:17 | |
| *** rongze has joined #heat | 06:17 | |
| *** pablosan_ has quit IRC | 06:20 | |
| *** blamar has joined #heat | 06:23 | |
| *** IlyaE has joined #heat | 06:26 | |
| *** tzumainn has quit IRC | 06:30 | |
| *** IlyaE has quit IRC | 06:32 | |
| *** nosnos_ has joined #heat | 06:43 | |
| *** nosnos has quit IRC | 06:43 | |
| *** tomek_adamczewsk has joined #heat | 06:54 | |
| shardy | morning all | 07:16 |
| *** aignatov is now known as aignatov_ | 07:20 | |
| skraynev | hello guys | 07:21 |
| skraynev | SpamapS: sorry, I read your message just now. | 07:22 |
| *** ramishra has quit IRC | 07:27 | |
| *** ramishra has joined #heat | 07:28 | |
| *** ramishra has quit IRC | 07:32 | |
| *** Linz has quit IRC | 07:38 | |
| *** Linz has joined #heat | 07:41 | |
| *** jistr has joined #heat | 07:43 | |
| *** aignatov_ is now known as aignatov | 07:54 | |
| *** e0ne has joined #heat | 07:54 | |
| *** ramishra has joined #heat | 07:56 | |
| openstackgerrit | Jenkins proposed a change to openstack/heat: Imported Translations from Transifex https://review.openstack.org/64504 | 07:59 |
| *** yogesh has joined #heat | 08:00 | |
| skraynev | shardy: have you a one minute? | 08:01 |
| sgran | hi all | 08:05 |
| sgran | I'm a little confused, I think | 08:05 |
| sgran | in the grizzly neutron/loadbalacner resource, it seems that you must supply an address to the vip, although the neutron api doesn't mandate this | 08:05 |
| *** jprovazn has joined #heat | 08:06 | |
| sgran | in HEAD, it looks like this might not be the case any more? I'm not sure, though | 08:06 |
| *** aignatov is now known as aignatov_ | 08:06 | |
| sgran | I was just going to start looking at how to make it an optional parameter, but it looks like it might already be optional - I wanted to check and see if anyone knows for sure | 08:06 |
| therve | sgran, I know for sure it's optional | 08:08 |
| shardy | skraynev: Hi, sure | 08:16 |
| skraynev | shardy: I only wanted ask you about how do we get updated list of requirements? | 08:17 |
| *** ramishra has quit IRC | 08:18 | |
| skraynev | shardy: I mean: Should it be automatically or somebody make patch with this changes (fetch changes from global-requirements)? | 08:18 |
| shardy | skraynev: You get the patch merged into openstack/requirements then there is a bot which automatically sends us a patch syncing us to the global requirements repo | 08:18 |
| *** renlt has quit IRC | 08:18 | |
| shardy | skraynev: previously it was a manual process, but AFAIK it's all automated now | 08:19 |
| skraynev | shardy: Well, I agree with you. I asked, because it's confused see bot's patch https://review.openstack.org/#/c/68240/ and other manual https://review.openstack.org/#/c/68017/ | 08:20 |
| skraynev | shardy: I understand, that it different changes, but I worry, that it;s right way. Should we review and approve manual patches now? | 08:22 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 08:23 | |
| shardy | skraynev: I would prefer we worked out why the automated sync is wrong | 08:23 |
| shardy | skraynev: It would not be the first time the bot got broken and the automated sync was wrong.. | 08:24 |
| *** tomek_adamczewsk has quit IRC | 08:24 | |
| shardy | but we should ping the infra guys and work out what is broken | 08:24 |
| shardy | skraynev: Your patch hasn't yet merged? | 08:27 |
| shardy | https://github.com/openstack/requirements/blob/master/global-requirements.txt | 08:27 |
| skraynev | shardy: not yet, it wait approve) | 08:27 |
| shardy | So it's impossible to merge the manual patch anyway, it should be abandoned and we should chase some folks to approve the requirements patch | 08:28 |
| sgran | therve: thank you :) That saved me some effort. Now to backport that ... | 08:31 |
| skraynev | shardy: well, when autosync will be to work correct, should we add -1 to manual update patch? | 08:31 |
| shardy | skraynev: I already have | 08:32 |
| skraynev | shardy: yeap) I have gotten your comments in ML notification) | 08:33 |
| therve | sgran, To grizzly? That might be a tough one | 08:34 |
| sgran | yes, I expect it to be :) | 08:34 |
| sgran | er, no, sorry | 08:34 |
| sgran | I was working against grizzly | 08:34 |
| sgran | I've upgraded my test setup to havana | 08:34 |
| sgran | so less so | 08:34 |
| sgran | sorry, more coffee needed :) | 08:34 |
| therve | You get a lot of other benefits in Havana, too | 08:34 |
| sgran | yeah, I'm tying up a few loose ends before wider deployment now | 08:35 |
| shardy | sgran: That would definitely be my suggestion, Havana unless you're really stuck with grizzly | 08:35 |
| sgran | so far I'm really happy | 08:35 |
| sgran | but it will still be a little work to backport the patch - I see you've rewritten the way schemas are validated | 08:35 |
| shardy | sgran: Have you figured out which bug # needs to be backported? | 08:37 |
| sgran | I don't think there is a bug number | 08:37 |
| sgran | I think that when the schema validation was redone, the status changed from required to optional | 08:37 |
| therve | Hum that would surprise me | 08:37 |
| shardy | Ok, or the patch which fixed it, we'll definitely need a bug and a reference to a commit in master to do a stable/havana backport | 08:38 |
| sgran | I'm looking at db4f855eb8c7675aec8510b0056056057925282a | 08:38 |
| shardy | sgran: Normally, you'd mark an existing fix committed bug as havana-backport-potential so the PTL will spot it | 08:38 |
| sgran | http://paste.debian.net/77737/ | 08:39 |
| sgran | I'll investigate today and see if I can come up with anything | 08:39 |
| sgran | I might also just live with a local patch. Icehouse isn't that far away | 08:39 |
| therve | sgran, Yeah that changes the syntax, but the meaning is the same | 08:40 |
| therve | It wasn't required at that time | 08:40 |
| sgran | ok | 08:40 |
| sgran | then let me figure out what's going wrong - it definitely doesn't work in havana without an address being passed. I'll see if I can see if it's just a bug. I thought that the absence of a default meant that it was required and stopped there | 08:40 |
| shardy | sgran: please don't live with a local patch if it's are real problem, report it as a bug so we can backport something & other users can benefit | 08:42 |
| shardy | s/are/a | 08:42 |
| *** cmyster has joined #heat | 08:44 | |
| sgran | sure. As I say, I thought it was just a schema issue that's now resolved. From what I'm hearing, it might be a real bug | 08:44 |
| sgran | I'll let you know later today | 08:44 |
| *** cmyster has quit IRC | 08:45 | |
| *** giulivo has joined #heat | 08:45 | |
| *** cmyster has joined #heat | 08:46 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Adding first_address attribute for port resource https://review.openstack.org/67836 | 08:51 |
| *** tsmadds has joined #heat | 08:55 | |
| *** derekh has joined #heat | 09:03 | |
| openstackgerrit | Zhang Yang proposed a change to openstack/heat: add the validation of MaxSize ,MinSize and DesiredCapacity https://review.openstack.org/67618 | 09:08 |
| *** aignatov_ is now known as aignatov | 09:10 | |
| sgran | ok, so, ignore me | 09:14 |
| sgran | it works just fine. I definitely have it on my TODO list as not working, but this might be from an earlier snapshot | 09:15 |
| sgran | but the version in havana is fine, so sorry to waste people's time :) | 09:15 |
| *** derekh has quit IRC | 09:18 | |
| *** mkollaro has joined #heat | 09:24 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 09:27 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 09:28 | |
| *** ramishra has joined #heat | 09:29 | |
| *** Linz has quit IRC | 09:32 | |
| *** ramishra has quit IRC | 09:33 | |
| *** ramishra has joined #heat | 09:34 | |
| *** aignatov is now known as aignatov_ | 09:34 | |
| *** Linz has joined #heat | 09:35 | |
| *** aignatov_ is now known as aignatov | 09:37 | |
| *** jistr has quit IRC | 09:37 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 09:40 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 09:47 | |
| *** aignatov is now known as aignatov_ | 09:49 | |
| *** aignatov_ is now known as aignatov | 09:53 | |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: A cloud-config resource based on SoftwareConfig https://review.openstack.org/63214 | 09:54 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: A multipart cloud-init resource based on SoftwareConfig https://review.openstack.org/63215 | 09:54 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: Resource type for software configuration https://review.openstack.org/67621 | 09:54 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: A cloud-config resource based on SoftwareConfig https://review.openstack.org/63214 | 09:56 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: A multipart cloud-init resource based on SoftwareConfig https://review.openstack.org/63215 | 09:56 |
| *** jistr has joined #heat | 09:57 | |
| *** alexpilotti has joined #heat | 09:58 | |
| liang | therve, Hi | 10:10 |
| therve | liang, Hello! | 10:10 |
| liang | therve, thanks for the review - https://review.openstack.org/#/c/61036/6 | 10:11 |
| liang | I just talk to Marconi people, and we agree that the version info should be taken care of by the client lib | 10:12 |
| *** akuznetsov has quit IRC | 10:12 | |
| therve | Yeah that all version in the endpoint thing is weird | 10:12 |
| liang | yeah | 10:13 |
| *** akuznetsov has joined #heat | 10:13 | |
| therve | liang, Presumably we can wait for the client change to land then? | 10:13 |
| liang | therve, well, it will make the endpoint configuration easier. But perhaps we can still proceed with the plugin patch? | 10:15 |
| liang | therve, BTW, I don't very get what you mean exactly by "directories are missing __init__.py files" | 10:16 |
| liang | I don't quite get ... | 10:17 |
| therve | Oh | 10:18 |
| *** yogesh has quit IRC | 10:18 | |
| therve | liang, forget it, it's just that they are empty so no created by the diff, sorry | 10:18 |
| therve | liang, I'm a bit reluctant to merge something that doesn't work out of the box with devstack. But I don't want to block your work either | 10:19 |
| liang | therve, no worries. I feel the same way actually, that why I made this attempt - https://review.openstack.org/#/c/60754 . Thanks for the review again ;) | 10:20 |
| *** derekh has joined #heat | 10:35 | |
| *** liang has quit IRC | 10:39 | |
| shardy | Anyone else having issues with the review interface jumping to the wrong location when you try to add a comment today? | 10:43 |
| cmyster | I had an issue with firefox crashing altogether | 10:47 |
| cmyster | in review that is | 10:48 |
| cmyster | but I'm not adding comments yet... | 10:49 |
| shardy | cmyster: I seem to have the opposite problem, firefox 26.0 works but Chrome 32.0.1700.77 doesn't | 10:50 |
| cmyster | idk, ff 24.0 kept crashing, chrome latest works fine but as I said I am not inserting comments yet | 10:51 |
| cmyster | shardy: pm ? | 10:56 |
| shardy | cmyster: ? | 10:57 |
| cmyster | just need a quick private question if you don't mind | 10:57 |
| shardy | cmyster: Ok, but in general you can just ask heat questions here | 10:58 |
| cmyster | I know | 10:58 |
| *** michchap has quit IRC | 11:06 | |
| *** michchap has joined #heat | 11:06 | |
| *** amritanshu_RnD has joined #heat | 11:08 | |
| *** gokrokve has quit IRC | 11:11 | |
| *** e0ne_ has joined #heat | 11:17 | |
| *** e0ne has quit IRC | 11:21 | |
| *** aignatov is now known as aignatov_ | 11:26 | |
| *** rongze has quit IRC | 11:26 | |
| *** rongze has joined #heat | 11:27 | |
| *** yogesh has joined #heat | 11:28 | |
| *** derekh has quit IRC | 11:29 | |
| *** rongze has quit IRC | 11:31 | |
| *** yogesh has quit IRC | 11:33 | |
| *** jeju has joined #heat | 11:34 | |
| jeju | hi | 11:34 |
| jeju | where can I get some heat templates to test autoscaling with ceilometer? | 11:35 |
| shardy | jeju: https://github.com/openstack/heat-templates/blob/master/cfn/F17/AutoScalingCeilometer.yaml | 11:36 |
| shardy | You may need to tweak it slightly to work with Fedora20 or some other distro | 11:36 |
| *** mkollaro has quit IRC | 11:36 | |
| jeju | shardy, thanks...but I forgot to mention that I don't want to use loadbalancer :| | 11:37 |
| shardy | jeju: You'll have to cut/paste your own template then | 11:38 |
| shardy | jeju: we welcome contributions if you find the existing examples don't contain what you need | 11:40 |
| shardy | Also note there's a minimal template example for every resource in the template guide: | 11:40 |
| shardy | http://docs.openstack.org/developer/heat/template_guide/index.html | 11:41 |
| jeju | shadower, actually i'm trying this template: http://pastebin.com/V7NYJDvR | 11:41 |
| jeju | ops | 11:41 |
| jeju | shardy, actually i'm trying this template: http://pastebin.com/V7NYJDvR | 11:41 |
| jeju | i'm using devstack, my stack is deployed correctly,but no alarm is triggered :( | 11:42 |
| shardy | jeju: I'd use the ceilometer CLI tool to check the alarms have been correctly created, then check the heat-engine logs for errors around the alarm notification | 11:44 |
| cmyster | is it not being triggered or not being sent/received | 11:44 |
| shardy | Sounds like probably a ceilometer issue | 11:45 |
| jeju | shardy, mmm alarm-list returns an empty table :| | 11:48 |
| shardy | jeju: but the ceilometer alarm resources are CREATE_COMPLETE? | 11:48 |
| shardy | heat resource-list <stack> | 11:48 |
| jeju | shardy, CPUlarmHigh | OS::Ceilometer::Alarm | CREATE_COMPLETE | 2014-01-22T11:49:08Z | 11:50 |
| *** aignatov_ is now known as aignatov | 11:50 | |
| jeju | shardy, mmm maybe I spotted a problem http://pastebin.com/nWyUVYuW | 11:53 |
| jeju | insufficient data!? | 11:53 |
| shardy | It means the alarm has insufficient metric datapoints to evaluate against the threshold | 11:54 |
| shardy | maybe the ceilometer agent collecting cpu metrics is not working or enabled? | 11:54 |
| jeju | shardy, in the ceilometer-acompute screen I see: | 11:55 |
| jeju | 2014-01-22 12:53:52.000 17335 AUDIT ceilometer.pipeline [-] Flush pipeline cpu_pipeline | 11:55 |
| jeju | so I think is working... | 11:56 |
| jeju | how can I check? | 11:56 |
| shardy | jeju: some combination of ceilometer meter-list and sample-list I guess | 11:57 |
| * shardy is not a ceilometer expert | 11:58 | |
| shardy | there are folks over in #openstack-metering who I'm sure can help | 11:58 |
| jeju | shardy, thank you :) | 11:59 |
| *** gokrokve has joined #heat | 12:06 | |
| *** nkhare has quit IRC | 12:06 | |
| *** ramishra_ has joined #heat | 12:09 | |
| *** rongze has joined #heat | 12:09 | |
| *** ramishra has quit IRC | 12:11 | |
| jeju | another question... have I to use an image with the cnftools to works with that template? Like F17-x86_64-cfntools or can I use a normal cloud image https://fedoraproject.org/en/get-fedora#clouds ? | 12:12 |
| *** gokrokve has quit IRC | 12:12 | |
| *** sergmelikyan has quit IRC | 12:12 | |
| shardy | jeju: From Fedora20 the normal cloud image contains heat-cfntools, so you can just use that | 12:13 |
| jeju | shardy, thanks :) | 12:13 |
| *** tspatzier has joined #heat | 12:13 | |
| shardy | jeju: There are pre-built older images here: http://fedorapeople.org/groups/heat/prebuilt-jeos-images/ | 12:13 |
| *** rongze has quit IRC | 12:14 | |
| *** dims has joined #heat | 12:14 | |
| jeju | mmm...can I manually trigger an alarm for test pourposes? | 12:15 |
| *** ifarkas has quit IRC | 12:16 | |
| *** bgorski has joined #heat | 12:17 | |
| *** ifarkas has joined #heat | 12:17 | |
| shardy | jeju: Yes, try ceilometer alarm-state-set | 12:19 |
| *** asalkeld has quit IRC | 12:19 | |
| shardy | also you can curl -X POST the URL of the ScalingPolicy directly, removing ceilometer from the loop completely | 12:19 |
| jeju | shardy, thanks!! | 12:21 |
| *** alexpilotti has quit IRC | 12:22 | |
| *** aignatov is now known as aignatov_ | 12:26 | |
| jeju | bye! | 12:31 |
| *** jeju has quit IRC | 12:31 | |
| *** rongze has joined #heat | 12:39 | |
| *** rpothier has quit IRC | 12:53 | |
| *** SnowDust has joined #heat | 12:53 | |
| *** scollier has joined #heat | 12:57 | |
| openstackgerrit | Thomas Herve proposed a change to openstack/heat: Don't pass empty security groups in port creation https://review.openstack.org/68073 | 12:58 |
| *** gokrokve has joined #heat | 13:01 | |
| *** gokrokve has quit IRC | 13:06 | |
| sdake | morning | 13:07 |
| shardy | Hi sdake | 13:08 |
| *** ramishra_ has quit IRC | 13:10 | |
| *** aignatov_ is now known as aignatov | 13:11 | |
| *** tsmadds has quit IRC | 13:19 | |
| *** radez_g0n3 is now known as radez | 13:21 | |
| *** tsmadds has joined #heat | 13:21 | |
| *** tspatzier has quit IRC | 13:24 | |
| *** tspatzier has joined #heat | 13:25 | |
| *** aweiteka has joined #heat | 13:29 | |
| *** tspatzier has quit IRC | 13:35 | |
| *** gilliard has joined #heat | 13:35 | |
| *** tspatzier has joined #heat | 13:35 | |
| *** derekh has joined #heat | 13:35 | |
| *** rongze has quit IRC | 13:37 | |
| *** blomquisg has quit IRC | 13:38 | |
| *** rpothier has joined #heat | 13:39 | |
| *** jdob has joined #heat | 13:39 | |
| *** achampion has quit IRC | 13:39 | |
| *** pafuent has joined #heat | 13:43 | |
| *** rongze has joined #heat | 13:47 | |
| *** SnowDust has quit IRC | 13:49 | |
| *** akuznetsov has quit IRC | 13:49 | |
| *** tnurlygayanov_ has joined #heat | 13:51 | |
| *** tnurlygayanov has quit IRC | 13:51 | |
| *** aignatov is now known as aignatov_ | 13:57 | |
| *** mestery has joined #heat | 13:57 | |
| *** tzumainn has joined #heat | 14:01 | |
| *** gokrokve has joined #heat | 14:02 | |
| *** akuznetsov has joined #heat | 14:04 | |
| *** gokrokve has quit IRC | 14:06 | |
| *** jasond` has joined #heat | 14:07 | |
| *** aignatov_ is now known as aignatov | 14:09 | |
| *** mkollaro has joined #heat | 14:10 | |
| *** ramishra has joined #heat | 14:11 | |
| *** tspatzier has quit IRC | 14:11 | |
| *** matsuhashi has joined #heat | 14:14 | |
| *** rongze_ has joined #heat | 14:14 | |
| sdake | http://imgur.com/n7l6n8C | 14:15 |
| *** ramishra has quit IRC | 14:16 | |
| *** rongze has quit IRC | 14:16 | |
| cmyster | mornin sdake | 14:17 |
| *** rongze has joined #heat | 14:17 | |
| *** rongze_ has quit IRC | 14:19 | |
| *** Linz has quit IRC | 14:19 | |
| *** Linz has joined #heat | 14:19 | |
| *** jcru has joined #heat | 14:20 | |
| sdake | hey cmyster | 14:20 |
| cmyster | I'm going over all the docs I can find. will start working with you soon enough ;) | 14:20 |
| radix | shardy: replied to your review. basically, there aren't two resources we're comparing - just a resource and a template | 14:21 |
| radix | (also, good morning heaters) | 14:21 |
| *** vijendar has joined #heat | 14:21 | |
| shardy | radix: Yeah, I was wondering if we should just create the resource objects to compare | 14:22 |
| shardy | maybe that's too much overhead though | 14:22 |
| *** rongze_ has joined #heat | 14:22 | |
| radix | you mean by applying the update? that would start disabling the instances before we have a chance to remove them from the LB | 14:22 |
| radix | which is the purpose of this function, to determine which instances to remove from the LB temporarily during an update | 14:22 |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Use a HeatIdentifier instance to create Parameters https://review.openstack.org/68257 | 14:23 |
| shardy | No, I just mean use the template to create an object (not to actually create anything, just a transient resource object which we discard) | 14:23 |
| shardy | Maybe it's a dumb idea, it was a pre-coffee review :) | 14:23 |
| radix | hehe :) well, I guess that would work, but it does seem like unnecessary LoC | 14:24 |
| radix | I'm not super happy about converting them to a string fwiw, but unfortunately frozendict doesn't exist :) | 14:25 |
| *** rongze has quit IRC | 14:25 | |
| *** arbylee has joined #heat | 14:27 | |
| pscheie | I'm trying to spin up a stack but getting a 'quota exceeded' error, says I've already used 998912 of 1000000 | 14:31 |
| pscheie | But in nova quota-show, it says the quota is 5000000 (because I just set it to that) | 14:31 |
| *** alexpilotti has joined #heat | 14:31 | |
| pscheie | Is there some place else quotas on memory are imposed? | 14:32 |
| *** ramishra has joined #heat | 14:32 | |
| shardy | zaneb: around? | 14:35 |
| *** andersonvom has joined #heat | 14:41 | |
| *** Linz has quit IRC | 14:43 | |
| *** arbylee has quit IRC | 14:43 | |
| *** arbylee has joined #heat | 14:44 | |
| *** matsuhashi has quit IRC | 14:45 | |
| *** pablosan has joined #heat | 14:48 | |
| *** rongze_ has quit IRC | 14:49 | |
| pscheie | So, it appears that while I've raised the RAM quota for a specific tenant, it's the default quota limit that I'm hitting for some reason. | 14:49 |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Native Pseudo Parameters https://review.openstack.org/66640 | 14:50 |
| pscheie | Even though I'm exporting the OS_TENANT_NAME env variable when calling heat. | 14:50 |
| openstackgerrit | Swann Croiset proposed a change to openstack/heat: Add autoscaling notifications https://review.openstack.org/62087 | 14:53 |
| scroiset | radix, stevebaker : ^^^^ should address your concerns. Let me know | 14:54 |
| *** tims has joined #heat | 14:56 | |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Prevent access Parameters key in template dict https://review.openstack.org/68312 | 14:56 |
| *** bvandenh has quit IRC | 14:57 | |
| *** nosnos_ has quit IRC | 14:57 | |
| *** tims1 has joined #heat | 14:58 | |
| *** tspatzier has joined #heat | 14:59 | |
| *** arbylee1 has joined #heat | 15:00 | |
| *** tims has quit IRC | 15:00 | |
| *** arbylee1 has quit IRC | 15:00 | |
| *** arbylee1 has joined #heat | 15:01 | |
| *** alexpilotti has quit IRC | 15:01 | |
| *** gokrokve has joined #heat | 15:03 | |
| *** arbylee has quit IRC | 15:04 | |
| *** rongze has joined #heat | 15:05 | |
| *** dims has quit IRC | 15:08 | |
| *** gokrokve has quit IRC | 15:08 | |
| *** dims has joined #heat | 15:09 | |
| *** gokrokve has joined #heat | 15:09 | |
| radix | hmm | 15:12 |
| radix | I guess heat has lots of code in __init__.py files | 15:12 |
| *** alexpilotti has joined #heat | 15:15 | |
| *** achampion has joined #heat | 15:20 | |
| *** rongze has quit IRC | 15:22 | |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add API support for stack preview https://review.openstack.org/68406 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add StackResource specific preview behavior https://review.openstack.org/68407 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add preview for NestedStack https://review.openstack.org/68408 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add preview for ResourceGroup https://review.openstack.org/68409 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add preview for InstanceGroup https://review.openstack.org/68410 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add preview for TemplateResource https://review.openstack.org/68411 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add preview for LoadBalancer https://review.openstack.org/68412 | 15:27 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add engine preview stack with Template and Params https://review.openstack.org/65576 | 15:27 |
| *** blomquisg has joined #heat | 15:28 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 15:33 | |
| *** amritanshu_RnD has quit IRC | 15:43 | |
| *** akuznetsov has quit IRC | 15:47 | |
| *** cmyster has quit IRC | 15:47 | |
| *** aignatov is now known as aignatov_ | 15:49 | |
| *** cmyster has joined #heat | 15:51 | |
| therve | God neutron | 15:51 |
| therve | Why are you so awful | 15:52 |
| *** akuznetsov has joined #heat | 15:55 | |
| *** jprovazn has quit IRC | 15:58 | |
| *** chandankumar_ has quit IRC | 15:58 | |
| radix | heh | 15:59 |
| *** nkhare has joined #heat | 16:00 | |
| *** achampio1 has joined #heat | 16:00 | |
| *** achampion has quit IRC | 16:00 | |
| *** chandankumar_ has joined #heat | 16:01 | |
| *** sballe has joined #heat | 16:02 | |
| *** jamieh has joined #heat | 16:02 | |
| *** rcleere has joined #heat | 16:02 | |
| *** achampio1 has quit IRC | 16:04 | |
| *** achampion has joined #heat | 16:05 | |
| *** gokrokve has quit IRC | 16:05 | |
| *** andersonvom1 has joined #heat | 16:06 | |
| *** andersonvom is now known as Guest63286 | 16:06 | |
| *** andersonvom1 is now known as andersonvom | 16:06 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_a | 16:07 | |
| *** Guest63286 has quit IRC | 16:07 | |
| *** SergeyLukjanov_a is now known as SergeyLukjanov_ | 16:08 | |
| *** achampion has quit IRC | 16:09 | |
| *** achampion has joined #heat | 16:10 | |
| zaneb | shardy: pong | 16:10 |
| shardy | zaneb: Hey, quick question about the new update code.. | 16:10 |
| shardy | whereabouts do we compare the old/new resource to see if it needs updating? | 16:11 |
| zaneb | it moved recently | 16:11 |
| shardy | I couldn't spot it post-refactor | 16:11 |
| zaneb | I think it's now in the resource itself | 16:11 |
| *** arbylee has joined #heat | 16:12 | |
| zaneb | https://github.com/openstack/heat/blob/master/heat/engine/resource.py#L507 | 16:12 |
| *** arbylee has quit IRC | 16:12 | |
| *** arbylee has joined #heat | 16:12 | |
| zaneb | https://github.com/openstack/heat/commit/c22d7e317af9fbc1140fed0fa496cdb791dd76f1 | 16:12 |
| zaneb | shardy: ^ | 16:12 |
| shardy | zaneb: Aha, I think I even reviewed that patch then completely forgot about it :) | 16:13 |
| shardy | zaneb: thanks | 16:13 |
| zaneb | np :) | 16:13 |
| bgorski | Hi all | 16:13 |
| shardy | zaneb: This is ref my comment on radix's patch https://review.openstack.org/#/c/68311/ | 16:13 |
| shardy | zaneb: I was trying to figure out if there's a cleaner, more update-ish way to do the comparison there | 16:13 |
| bgorski | I was off for awhile. | 16:14 |
| shardy | bgorski: o/ | 16:14 |
| *** kebray_ has joined #heat | 16:14 | |
| *** pafuent has quit IRC | 16:14 | |
| bgorski | I have a question about changing the requirements | 16:14 |
| zaneb | shardy: there is, but not for now | 16:14 |
| bgorski | I need to bump up the keystoneclient version to the 0.4.2 | 16:14 |
| *** arbylee1 has quit IRC | 16:15 | |
| bgorski | What is the right way to do that? | 16:15 |
| zaneb | shardy: as I was saying to radix last night, what I want to do is to implement a ResourceTemplate class. That will help us abstract the differences between HOT and cfn templates, and we can also make it hashable | 16:15 |
| shardy | bgorski: submit a patch to openstack/requirements, then a bot will propose a change to heat syncing after it gets merged | 16:15 |
| shardy | zaneb: Ok, sounds good - I couldn't get away from the feeling that comparing sets of json strings wasn't the way to go long term | 16:16 |
| zaneb | shardy: agreed, long term there are better ideas. short term this is pretty good | 16:17 |
| shardy | but I didn't have any fully-formed idea of a good alternative :) | 16:17 |
| zaneb | I'll leave a comment to that effect | 16:17 |
| shardy | zaneb: Ok, I'll remove my objection on the patch - if you could comment re the vision for future refactoring that would be helpful, thanks | 16:18 |
| shardy | bgorski: https://github.com/openstack/requirements | 16:19 |
| *** aignatov_ is now known as aignatov | 16:19 | |
| *** _ruhe is now known as ruhe | 16:20 | |
| *** denis_makogon has quit IRC | 16:20 | |
| radix | zaneb, shardy: shall I add a comment to the code? | 16:22 |
| radix | maybe we should have a placeholder blueprint to point at? | 16:22 |
| zaneb | tbh I don't think a comment is necessary | 16:23 |
| radix | okie doke | 16:23 |
| zaneb | a blueprint is a good idea though | 16:23 |
| zaneb | I'll raise one | 16:23 |
| shardy | +1 blueprint and a comment on the review should be enough | 16:24 |
| *** denis_makogon has joined #heat | 16:25 | |
| bgorski | shardy, Thx for info I will do that | 16:26 |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 16:28 | |
| bgorski | shardy, the patch with the change I need is already in progress https://review.openstack.org/#/c/66494/ | 16:32 |
| *** ramishra has quit IRC | 16:32 | |
| *** radez has quit IRC | 16:33 | |
| *** jomara_ has quit IRC | 16:33 | |
| shardy | bgorski: all you can do is +1 it and wait then | 16:33 |
| *** sgordon` has quit IRC | 16:33 | |
| *** aweiteka has quit IRC | 16:34 | |
| *** blomquisg has quit IRC | 16:34 | |
| *** pafuent has joined #heat | 16:36 | |
| radix | zaneb: are you suggesting serialize_template be just a local in the changing_instances function, or up at the module level? | 16:40 |
| zaneb | radix: just a local is fine | 16:40 |
| radix | ok | 16:40 |
| zaneb | or module level is fine | 16:40 |
| zaneb | either is fine ;) | 16:40 |
| zaneb | but I was assuming local | 16:40 |
| *** nkhare has quit IRC | 16:41 | |
| *** ramishra has joined #heat | 16:42 | |
| openstackgerrit | Christopher Armstrong proposed a change to openstack/heat: Make LB-updating in rolling update more reliable https://review.openstack.org/68311 | 16:42 |
| *** aignatov is now known as aignatov_ | 16:42 | |
| *** e0ne_ has quit IRC | 16:43 | |
| *** mkollaro has quit IRC | 16:44 | |
| zaneb | radix, shardy: https://blueprints.launchpad.net/heat/+spec/resource-template-api | 16:44 |
| *** radez has joined #heat | 16:44 | |
| *** jomara has joined #heat | 16:45 | |
| *** aweiteka has joined #heat | 16:45 | |
| openstackgerrit | Randall Burt proposed a change to openstack/heat: Refactor CLB to work with groups https://review.openstack.org/65586 | 16:45 |
| *** cmyster has quit IRC | 16:45 | |
| *** ramishra has quit IRC | 16:46 | |
| *** blomquisg has joined #heat | 16:47 | |
| *** funzo has quit IRC | 16:47 | |
| *** funzo has joined #heat | 16:47 | |
| *** randallburt has joined #heat | 16:53 | |
| *** jomara has quit IRC | 16:55 | |
| *** aweiteka has quit IRC | 16:55 | |
| *** radez has quit IRC | 16:55 | |
| *** john-n-seattle1 has joined #heat | 16:56 | |
| *** blomquisg has quit IRC | 16:57 | |
| pafuent | jasond: jasond`: Are you around? | 17:04 |
| jasond` | pafuent: yes | 17:05 |
| jasond` | pafuent: was just about to reply to you | 17:05 |
| zaneb | pafuent: thanks for that comment on the function-plugins blueprint, I completely agree | 17:05 |
| *** radez has joined #heat | 17:06 | |
| openstackgerrit | Bartosz Górski proposed a change to openstack/heat: Adding option in config to specify region name https://review.openstack.org/43335 | 17:07 |
| pafuent | jasond`: Ok. Thanks. I was looking for you regarding to that. Sorry for the anxiety ;) | 17:07 |
| *** Linz has joined #heat | 17:07 | |
| *** aweiteka has joined #heat | 17:08 | |
| jasond` | pafuent: no problem. i just replied | 17:08 |
| pafuent | zaneb: You are welcome | 17:09 |
| *** jomara_ has joined #heat | 17:10 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 17:10 | |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: HOT templates get_attr allows extra attributes https://review.openstack.org/67562 | 17:14 |
| *** pshchelo has quit IRC | 17:15 | |
| *** aweiteka has quit IRC | 17:15 | |
| *** jomara_ has quit IRC | 17:15 | |
| *** radez has quit IRC | 17:16 | |
| *** bgorski has quit IRC | 17:18 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 17:18 | |
| *** jomara_ has joined #heat | 17:20 | |
| *** Linz has quit IRC | 17:21 | |
| *** radez has joined #heat | 17:22 | |
| *** chandankumar_ has quit IRC | 17:24 | |
| *** aweiteka has joined #heat | 17:25 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 17:28 | |
| *** sjmc7 has joined #heat | 17:28 | |
| *** pvaneck has joined #heat | 17:28 | |
| *** chandankumar_ has joined #heat | 17:29 | |
| *** jmckind has joined #heat | 17:30 | |
| *** jistr has quit IRC | 17:32 | |
| randallburt | andersonvom: could you update https://blueprints.launchpad.net/heat/+spec/preview-stack with example output? I'm having a hard time understanding the "why" of https://review.openstack.org/#/c/68412 | 17:32 |
| *** blamar_ has joined #heat | 17:40 | |
| andersonvom | randallburt: I'll take a look at it. but that was really just because LoadBalancer is technically a StackResource | 17:43 |
| *** blamar has quit IRC | 17:43 | |
| *** blamar_ is now known as blamar | 17:43 | |
| randallburt | andersonvom: thanks | 17:43 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add preview for LoadBalancer https://review.openstack.org/68412 | 17:46 |
| andersonvom | randallburt: changed the commit message to make that clear | 17:46 |
| *** julienvey has left #heat | 17:46 | |
| andersonvom | but I'll look into the bp as well | 17:47 |
| randallburt | andersonvom: cool. that part I kinda got, its just the looking for the KeyPair stuff that confused me. | 17:47 |
| andersonvom | randallburt: yeah, since it's a stackresource, it needs a template to be parsed, and KeyName is required, so it needed that as well | 17:48 |
| randallburt | andersonvom: so if that's a required param of the nested template, It should probably be documented in the resource I would think. I commented on the review. A weak −1 for now until i can see some example output and a response to the comment on the review. Thanks for the info! | 17:53 |
| randallburt | actually, andersonvom, looking at the old code, it looks like it was making that assumption all along. Still think it needs to be documented in some way, but that's probably not on you IMO. | 17:54 |
| shardy | zaneb: Another question about the new update logic if you have a sec? | 17:54 |
| zaneb | go ahead | 17:55 |
| shardy | zaneb: I'm wondering what the expected behavior is re the backup stack - does update really create a new stack and flip the ids, then delete the old one? | 17:56 |
| zaneb | no | 17:57 |
| shardy | zaneb: I'm seeing the backup stack trigger deletion of the trust created by the initial stack, so trying to work out the expected update order before digging deeper | 17:57 |
| zaneb | it creates the backup stack, moves resources around between them, then deletes the backup stack | 17:57 |
| shardy | Ah, but the backup stack is a copy of everything, including the user_creds record? | 17:58 |
| * shardy looks | 17:58 | |
| zaneb | not sure | 17:58 |
| zaneb | but sounds like it would explain your issue | 17:58 |
| shardy | zaneb: No worries, I can dig myself, I just wanted to make sure I didn't have any fundamental misunderstanding of how it's supposed to work :) | 17:59 |
| shardy | zaneb: Yeah that is the problem, thanks | 18:03 |
| zaneb | cool | 18:03 |
| zaneb | hopefully that's relatively simple to fix | 18:04 |
| shardy | yeah, just pass a backup argument to delete and don't delete the trust | 18:04 |
| *** saurabhs has joined #heat | 18:06 | |
| *** jasond` has quit IRC | 18:06 | |
| zaneb | ffs | 18:07 |
| zaneb | I ordered a monitor and keyboard on Monday | 18:07 |
| zaneb | for 3 days they have been asking questions about random businesses on streets that happen to contain some of the same words as my street | 18:08 |
| zaneb | today they inform me that I got the wrong part number for "this item", and that I can't buy it because I have no upgrades. Instead I need to buy a $400 software license to some thing I have never heard of | 18:09 |
| zaneb | I ask "which item?" and they reply "the one you ordered" | 18:09 |
| *** yogesh has joined #heat | 18:12 | |
| zaneb | sdake: ^ early status report for the week :/ | 18:12 |
| *** derekh has quit IRC | 18:16 | |
| *** gokrokve has joined #heat | 18:17 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 18:21 | |
| openstackgerrit | Thomas Spatzier proposed a change to openstack/heat: Enable better sub-classing of common Schema class https://review.openstack.org/65688 | 18:25 |
| openstackgerrit | Thomas Spatzier proposed a change to openstack/heat: Refactor Parameters Schema based on common Schema https://review.openstack.org/67183 | 18:25 |
| openstackgerrit | Thomas Spatzier proposed a change to openstack/heat: Move param format for template-validate to API https://review.openstack.org/67171 | 18:25 |
| *** tsmadds has quit IRC | 18:32 | |
| *** harlowja_away is now known as harlowja | 18:32 | |
| *** tspatzier has quit IRC | 18:36 | |
| *** arbylee has quit IRC | 18:37 | |
| *** ruhe is now known as _ruhe | 18:39 | |
| *** jamieh has quit IRC | 18:42 | |
| *** jprovazn has joined #heat | 18:50 | |
| stevebaker | morning | 18:51 |
| shardy | hi stevebaker | 18:51 |
| stevebaker | shardy: hi | 18:51 |
| *** e0ne has joined #heat | 18:52 | |
| *** kfox1111 has joined #heat | 18:55 | |
| kfox1111 | Anyone know why I might get this on stack create: Resource create failed: StackValidationFailed: Unknown resource Type : AWS::CloudWatch::Alarm | 18:56 |
| *** andrew_plunk has joined #heat | 18:56 | |
| kfox1111 | when I don't have an alarm defined in the stack? | 18:56 |
| *** blomquisg has joined #heat | 18:56 | |
| SpamapS | how did we not stop i2 with this one: https://bugs.launchpad.net/heat/+bug/1271190 ? | 18:57 |
| stevebaker | hmm, I think the alt metting time is 0000 UTC on *Thursday* not Wednesday (unless I missed it?) https://wiki.openstack.org/wiki/Meetings/HeatAgenda#Weekly_Heat_.28Orchestration.29_meeting | 18:57 |
| SpamapS | Heat i2 is basically unusable for any use case that needs in-instance users. | 18:57 |
| shardy | SpamapS: I marked it critical and targetted it at I-2... | 18:58 |
| SpamapS | shardy: yeah, Thierry just moved it and released | 18:58 |
| SpamapS | which is.. | 18:58 |
| SpamapS | basically broken | 18:58 |
| shardy | stevebaker: Did he run that past you? | 18:58 |
| SpamapS | I mean, I don't really believe in the release milestones as anything other than "get your crap together so we can show the world what we're up to".. but at this point.. i2 is "we're up to breaking things" | 18:59 |
| *** tango has joined #heat | 18:59 | |
| stevebaker | A call was made project-wide to just cut i-2 due to the gate queue length. I don't know if there is opportunity for backports | 18:59 |
| SpamapS | _ah_ | 18:59 |
| stevebaker | but I did raise that bug explicitly as needed | 18:59 |
| SpamapS | so basically everybody is broken :) | 19:00 |
| *** nati_ueno has joined #heat | 19:00 | |
| SpamapS | well as long as its _equal_ | 19:00 |
| stevebaker | yes, really not ideal | 19:00 |
| shardy | stevebaker: Ok, sounds really dumb, but whatever - I would think delaying it by a week would be better than releasing a bunch of stuff folks can't even test | 19:00 |
| *** _ruhe is now known as ruhe | 19:01 | |
| stevebaker | shardy: that was an option, but that eats into i-3 which arguably deserves the extra time | 19:01 |
| stevebaker | I'm going to update the wiki to make the alt time to be Thursday 0000 UTC | 19:03 |
| shardy | stevebaker: Well the real thing blocking everyone for progress on I-3 is the gate not the milestone | 19:03 |
| SpamapS | This is a never ending argument | 19:03 |
| SpamapS | it is "march of death" vs. "rest and be ready" | 19:03 |
| shardy | SpamapS: well it's frustrating when the fixes are all ready and the issue is the gate | 19:04 |
| shardy | Oh well, nevermind | 19:04 |
| stevebaker | is it still in the gate? | 19:04 |
| SpamapS | shardy: yeah, if only we could attack the gate bugs with more people. :-P | 19:04 |
| SpamapS | stevebaker: queued | 19:04 |
| shardy | stevebaker: https://review.openstack.org/#/c/68135/ | 19:05 |
| *** arbylee has joined #heat | 19:06 | |
| stevebaker | Its not only gate bugs, it is also that review load is exceeding available cloud resources. Wheels are turning to increase resources. The bugs do compound that though | 19:06 |
| zaneb | AIUI ttx believes the that i-1/i-2/i-3 releases don't mean anything, and are just dress-rehearsals for the RCs | 19:10 |
| zaneb | personally, I don't understand why they exist if that is the case | 19:11 |
| zaneb | and things like RDO absolutely do package them | 19:11 |
| zaneb | and it is a giant PITA if they are broken and no backports are allowed | 19:11 |
| shardy | zaneb: +1 | 19:13 |
| zaneb | stevebaker: I haven't been to one of the release meetings in this cycle. Have they gotten any better? Or still just pointless busywork? | 19:14 |
| kfox1111 | Is it a requirement to have alarm resources even if you are not using them? | 19:16 |
| kfox1111 | zaneb: agreed. I might have tried i2, but not now. :/ | 19:17 |
| stevebaker | zaneb: they are a bit better, there is more of a focus on cross-project blocking issues now that ttx does a 1-1 for the release bookeeping stuff | 19:18 |
| *** arbylee has quit IRC | 19:18 | |
| *** arbylee has joined #heat | 19:18 | |
| zaneb | stevebaker: what about stuff like 'feature x in project y is needed in project z, so we will use release icehouse-n as a synchronisation point'? | 19:19 |
| zaneb | that's what should be happening in those meetings IMO | 19:20 |
| pafuent | jasond: I can't reproduce your comment. I ran the test with your change and the tests pass. Maybe I'm missing something. | 19:20 |
| *** edmund has joined #heat | 19:20 | |
| zaneb | stevebaker: e.g. we renamed quantum to neutron, so everybody update your client lib references before H-2 | 19:20 |
| zaneb | something that never happened but should have ^ | 19:21 |
| stevebaker | zaneb: that happens more. not that we've had anything that disruptive in this cycle | 19:21 |
| zaneb | kfox1111: there's no conceivable way that it Heat could be complaining about a particular type of resource you don't have in your template | 19:22 |
| stevebaker | recently it has been mostly brainstorming on how to improve the gate | 19:22 |
| zaneb | stevebaker: ok, that's a good sign, thanks | 19:22 |
| *** e0ne_ has joined #heat | 19:22 | |
| *** jprovazn has quit IRC | 19:22 | |
| zaneb | fix neutron? | 19:22 |
| zaneb | just throwing out ideas here | 19:22 |
| *** spzala has joined #heat | 19:23 | |
| stevebaker | To be fair, most of the recent problems have been nova bugs and load related timeouts due to test concurrency | 19:23 |
| kfox1111 | zaneb: That is what I was thinking, but then... I'm seeing it. :/ | 19:23 |
| kfox1111 | I'm turning up debugging logs to see if I can get more info. With debugging off, everything looks normal, just it fails with that. | 19:24 |
| zaneb | kfox1111: is this a template you can post publicly? paste.openstack.org | 19:24 |
| therve | kfox1111, Are you using the AWS LoadBalancer resource? | 19:24 |
| kfox1111 | zaneb: sure. give me a sec. Its basically just one of the heat-template examples I tweeked slightly and removed a few things to test. | 19:25 |
| kfox1111 | therve: yeah. | 19:25 |
| *** e0ne has quit IRC | 19:25 | |
| therve | kfox1111, There you go. It creates an alarm. | 19:25 |
| stevebaker | So heat i-2 tagging will be delayed until https://review.openstack.org/#/c/68135/ lands, so please keep an eye on it to nurse it through the gate | 19:26 |
| zaneb | therve: well spotted, thanks :) | 19:26 |
| kfox1111 | therve: even if you didn't define one? | 19:26 |
| zaneb | kfox1111: yes. the loadbalancer is implemented as a nested stack | 19:27 |
| therve | kfox1111, It does it for you. It needs one to monitor the instances. | 19:27 |
| zaneb | the nested stack contains an Alarm | 19:27 |
| kfox1111 | Ah. ok. | 19:28 |
| kfox1111 | what does it try and alarm on? | 19:28 |
| *** tsmadds has joined #heat | 19:28 | |
| zaneb | kfox1111: latency. I'm not even sure why. | 19:31 |
| *** topol has joined #heat | 19:32 | |
| *** tsmadds has quit IRC | 19:33 | |
| andersonvom | hey folks, is there supposed to be any sort of folder structure in the contrib/ folder? i.e. does heat expect any specific folder structure? | 19:33 |
| kfox1111 | zaneb: hmm... ok. thanks. | 19:37 |
| kfox1111 | So, in /etc/heat/environment.d/default.yaml I have: "AWS::CloudWatch::Alarm": "OS::Heat::CWLiteAlarm" and its still complaining about not knowing about that resource type. Is that an unsupported resource in havana? | 19:39 |
| *** randallburt has quit IRC | 19:47 | |
| radix | there's no heat meeting in 13 minutes, right? | 19:47 |
| radix | this is alternate-meeting week? | 19:47 |
| sdake | zaneb software for a monitor, what will they think up next :) | 19:48 |
| stevebaker | radix: its alt meeting time | 19:49 |
| radix | okie doke | 19:49 |
| stevebaker | 4hours to go | 19:49 |
| zaneb | sdake: I'll tell you want they won't think up: "Yes, we've solved the problem and we'll deliver it overnight for free." | 19:49 |
| sdake | zaneb lol | 19:50 |
| zaneb | only "we're sorry for the inconvenience" | 19:51 |
| zaneb | bull***t you are | 19:52 |
| *** tspatzier has joined #heat | 19:57 | |
| *** Qin__ has joined #heat | 19:59 | |
| stevebaker | PSA the heat meeting will be at the alternate time of 0000 UTC | 20:02 |
| *** randallburt has joined #heat | 20:03 | |
| randallburt | andersonvom: nope, not at the moment | 20:07 |
| randallburt | andersonvom: are you working on the test results stuff? | 20:07 |
| *** Qin__ has quit IRC | 20:09 | |
| *** derekh has joined #heat | 20:11 | |
| skraynev | Hello all, Could somebody else review https://review.openstack.org/#/c/65970/ . needs other opinion on question in comment. | 20:11 |
| andersonvom | randallburt: I'm trying to figure out how to generate the right docs for the contrib/ folder, so I was wondering if I can assume that plugins will always be inside contrib/**/heat/engine/plugins | 20:12 |
| randallburt | andersonvom: ah, I see. So no, there's no real restriction on the organization of contrib. | 20:12 |
| andersonvom | randallburt: yeah, looking at the folder right now, it doesn't. but I'm wondering if it should. | 20:13 |
| andersonvom | mostly to be able to scale the documentation generation | 20:13 |
| randallburt | andersonvom: can sphinx help you here? can you just point it at /contrib and say "go find things to document"? | 20:13 |
| randallburt | andersonvom: IIRC it was stevebaker that originally wrote the resource documentation generator, he may have some pointers. | 20:14 |
| andersonvom | randallburt: sort of. the thing is that we generate resource documentation separately from everything else, and when it imports heat, the configuration for plugins is the default value (/usr/lib/something) | 20:15 |
| randallburt | andersonvom: ah, yeah, contrib may need some organization then or perhaps that code needs updating to account for the plugins. maybe some of both. | 20:15 |
| andersonvom | randallburt: and then, the plugins are not loaded and left without docs | 20:15 |
| randallburt | wait, couldn't you just use the plugin loading mechanism? | 20:16 |
| stevebaker | it should be just a matter of adding contrib to the plugins path during docs generation | 20:16 |
| randallburt | ^^ | 20:16 |
| stevebaker | also if the resource requires a lib for it to register itself then that lib will need to be in the gated requirements | 20:17 |
| andersonvom | stevebaker: tried that and it does most of it. but the resources don't get any docs because at that point in time, the plugin folder is not correctly set up | 20:17 |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Use a HeatIdentifier instance to create Parameters https://review.openstack.org/68257 | 20:18 |
| stevebaker | also the resourcepages directive filters by root namespace | 20:18 |
| randallburt | plus, the plugins we do have check that a lib is available before they register, so they won't register themselves. That'll be a problem | 20:18 |
| stevebaker | andersonvom: what resources are you talking specifically? | 20:19 |
| andersonvom | stevebaker: so, the way I got it to work was loading the plugin_dirs config option and pointing that to contrib/rackspace/..../plugins | 20:19 |
| andersonvom | stevebaker: let me get a list for you | 20:19 |
| randallburt | andersonvom: don't forget the Docker one too, that's also missing from the published docs. | 20:19 |
| randallburt | at least it was the last time I bothered running the docs. | 20:19 |
| randallburt | yup. its missing too. there's a Rackspace section that's empty and no docker one at all. stevebaker might be worth changing that template and just use a "Contributed Plug-ins" section with some note saying they are in-tree but aren't loaded by default and may have other third-party library requirements. | 20:21 |
| stevebaker | randallburt: yep | 20:22 |
| randallburt | andersonvom: any of that help? | 20:22 |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Native Pseudo Parameters https://review.openstack.org/66640 | 20:22 |
| andersonvom | stevebaker: so, by adding contrib to the sources, I got the modules to show up in the docs, but when I look at template_guide, the http://10.5.72.146/build/html/template_guide/rackspace.html is empty, because no plugins were loaded, since cfg.CONF.plugin_dirs points to the default places, not contrib | 20:23 |
| stevebaker | can you set cfg.CONF.plugin_dirs in conf.py? | 20:23 |
| andersonvom | stevebaker: I can, and that's what led me to ask if contrib has any structure, because I'd like to add this in a way that new plugins get added to the docs automatically | 20:24 |
| stevebaker | I assume just adding contrib/ is enough. Any plugins will be found | 20:25 |
| randallburt | well even if they were loaded, they won't be registered since you will have missing dependencies. For example: https://github.com/openstack/heat/blob/master/contrib/docker-plugin/plugin/docker_container.py#L274-L282 | 20:25 |
| andersonvom | stevebaker: oh, I was under the assumption you had to point to the plugins folder specifically. let me test that | 20:25 |
| stevebaker | I don't think so | 20:26 |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 20:28 | |
| *** tsmadds has joined #heat | 20:29 | |
| andersonvom | stevebaker: cool. pointing to contrib solves it. I'll look into what randallburt said about the missing dependencies then | 20:30 |
| randallburt | andersonvom: well if it worked, then you can ignore my chicken little routine ;) | 20:30 |
| *** akuznetsov has quit IRC | 20:31 | |
| stevebaker | andersonvom: good. Also feel free to reorganise rackspace.rst to refer to all contrib resources | 20:33 |
| *** kebray_ has quit IRC | 20:34 | |
| randallburt | stevebaker: if its working, it might be easier/more maintainable to refactor that section like above | 20:34 |
| randallburt | since there's Docker in there now too. | 20:34 |
| randallburt | stevebaker: DOH! I re-read and realized that's probably what you meant. | 20:35 |
| stevebaker | thats ok, you can take the credit ;) | 20:35 |
| randallburt | no no, might just need more coffee ;) | 20:35 |
| *** tsmadds has quit IRC | 20:35 | |
| openstackgerrit | Pablo Andres Fuente proposed a change to openstack/heat: Prevent access Parameters key in template dict https://review.openstack.org/68312 | 20:36 |
| andersonvom | stevebaker: will do, thanks! | 20:38 |
| *** tspatzier has quit IRC | 20:43 | |
| *** rcleere has quit IRC | 20:46 | |
| *** kebray_ has joined #heat | 20:49 | |
| *** SergeyLukjanov_ is now known as SergeyLukjanov | 20:51 | |
| *** tsmadds has joined #heat | 20:55 | |
| *** tsmadds has quit IRC | 21:01 | |
| *** SergeyLukjanov is now known as SergeyLukjanov_ | 21:02 | |
| *** saurabhs has quit IRC | 21:06 | |
| *** saurabhs has joined #heat | 21:07 | |
| *** pafuent has left #heat | 21:10 | |
| *** kebray_ has quit IRC | 21:12 | |
| *** ruhe is now known as _ruhe | 21:23 | |
| *** tango|2 has joined #heat | 21:29 | |
| *** asalkeld has joined #heat | 21:30 | |
| *** tango has quit IRC | 21:32 | |
| larsks | Hello all. Should I be able to update the *outputs* of a stack without heat trying to recreate resources? | 21:44 |
| larsks | Huh, it blew away all my existing outputs, too. | 21:49 |
| *** tsmadds has joined #heat | 21:57 | |
| radix | so can someone remind me, if there's a stack with, say, a server in an error state, and I update-stack with an unchanged template, will the server be deleted and recreated? | 21:58 |
| stevebaker | zaneb: has behaviour for ^ changed at all yet? | 22:01 |
| zaneb | radix: no :( | 22:02 |
| *** tsmadds has quit IRC | 22:02 | |
| radix | oh, ok | 22:02 |
| cyli | zaneb: apologies, I'm really new to this so maybe this is a stupid question. Can we delete (just clean up) the errored/failed resources? | 22:03 |
| *** derekh has quit IRC | 22:03 | |
| zaneb | cyli: ideally, yes | 22:04 |
| radix | so you can remove the resources from the template, update-stack, and then re-add them to the template, and update-stack, to recreate the resources? | 22:04 |
| *** david-lyle has quit IRC | 22:06 | |
| zaneb | radix: you can't update a failed stack at all | 22:06 |
| radix | oh | 22:06 |
| *** david-lyle has joined #heat | 22:06 | |
| zaneb | delete or nothing | 22:06 |
| radix | right, okay. I was trying to find the code that might reject an update to a failed stack but couldn't | 22:06 |
| zaneb | as far as I recall, anyway | 22:07 |
| *** rpothier has quit IRC | 22:07 | |
| kfox1111 | Does heat without ceilometer in havana just not work with AWS::ElasticLoadBalancing::LoadBalancer? | 22:07 |
| radix | that is pretty damn bad from an autoscaling perspective :( | 22:07 |
| radix | I mean, it's pretty damn bad for any case, heh | 22:07 |
| zaneb | radix: https://github.com/openstack/heat/blob/master/heat/engine/parser.py#L492 | 22:08 |
| *** rpothier has joined #heat | 22:08 | |
| radix | zaneb: ah ok, thanks | 22:09 |
| radix | I wonder if the stack-converge bps will address this somehow | 22:11 |
| stevebaker | I would have thought that allowing an update on a FAILED stack would be a relatively small change, compared to a full converge | 22:13 |
| stevebaker | just flag all FAILED resources for UpdateReplace and do the Update | 22:13 |
| radix | that'd be pretty cool | 22:14 |
| radix | stevebaker: shall I create a bp for this? or maybe just a mailing list thread? | 22:17 |
| stevebaker | radix: bingo https://review.openstack.org/#/c/62936/2 | 22:18 |
| stevebaker | give SpamapS some review love | 22:18 |
| radix | oh my! | 22:19 |
| radix | yes sir :) | 22:19 |
| radix | this... changes... EVERYTHING | 22:19 |
| SpamapS | My retry-on-update stuff is stuck on me testing all the scenarios | 22:19 |
| SpamapS | I have a new patch which will be more intelligent | 22:20 |
| radix | ok cool | 22:20 |
| SpamapS | radix: it works reasonably well as it is.. | 22:20 |
| SpamapS | until it doesn't | 22:20 |
| radix | heh | 22:20 |
| SpamapS | and then .. sadness | 22:20 |
| radix | SpamapS: well, I'll put this on my watch-list :) | 22:21 |
| SpamapS | like if delete failed.. we resurrect it rather than trying the delete again ;) | 22:21 |
| SpamapS | it just involves actually caring about the action, not just the status | 22:21 |
| radix | ah, right. | 22:21 |
| radix | so now I'm getting curious about the check-stacu bp | 22:22 |
| radix | er, check-stack | 22:23 |
| radix | like, if it sets resource states to "FAILED", then check-stack + SpamapS-patch == convergence | 22:23 |
| radix | (with a cron job or something, which admittedly isn't great; some day we'll have queue-watching I'm sure...) | 22:23 |
| SpamapS | radix: \o/ thats exactly what I was thinking too :) | 22:24 |
| SpamapS | but.. shhhh | 22:24 |
| radix | heh heh :) | 22:24 |
| SpamapS | converge is such a cool command | 22:24 |
| SpamapS | it can't just be a chain of two others... can it? :) | 22:25 |
| radix | heat dwim | 22:25 |
| SpamapS | --force | 22:25 |
| *** jdob has quit IRC | 22:27 | |
| *** mkollaro has joined #heat | 22:28 | |
| *** aweiteka has quit IRC | 22:30 | |
| *** jergerber has joined #heat | 22:31 | |
| *** vijendar has quit IRC | 22:32 | |
| *** alexpilotti has quit IRC | 22:34 | |
| *** alexpilotti has joined #heat | 22:34 | |
| zaneb | radix: the bp already exists https://blueprints.launchpad.net/heat/+spec/update-failure-recovery | 22:35 |
| radix | yeah, I figured that out based on SpamapS's patch | 22:35 |
| *** bvandenh has joined #heat | 22:35 | |
| radix | oh wait, that's a different bp | 22:36 |
| sdake | going to step out for a bit - feeling a bit ill today with cold or something bbl | 22:37 |
| *** sdake is now known as sdake-ooo | 22:37 | |
| *** tango has joined #heat | 22:40 | |
| *** alexpilotti has quit IRC | 22:41 | |
| *** tango|2 has quit IRC | 22:44 | |
| *** achampion has quit IRC | 22:44 | |
| *** tango has quit IRC | 22:45 | |
| *** rpothier has quit IRC | 22:45 | |
| *** Winnie has joined #heat | 22:49 | |
| *** Winnie is now known as Guest76965 | 22:49 | |
| *** tango has joined #heat | 22:50 | |
| zaneb | SpamapS: I added some comments for you to think about | 22:52 |
| SpamapS | zaneb: many thanks | 22:58 |
| SpamapS | zaneb: we have to get to update-failed-recovery .. or we're lost | 22:58 |
| SpamapS | zaneb: we cannot delete a 1000 node cloud without good reason. :) | 22:58 |
| zaneb | I agree we have to get there | 22:58 |
| *** radez is now known as radez_g0n3 | 23:00 | |
| zaneb | I guess the question is do we release something that might work in some specific cases, or wait until we think we have a reasonable shot at supporting it in the general case | 23:00 |
| zaneb | I had hopes of working on that this cycle | 23:00 |
| zaneb | hasn't happened yet though :/ | 23:00 |
| lifeless | zaneb: why is it either-or? | 23:00 |
| SpamapS | zaneb: So if we extrapolate my approach for the general case.. a delete failed will just be attempted again. An update failed resource will just have update attempted again, and a create failed will be deleted/re-created. | 23:01 |
| zaneb | lifeless: SpamapS has a patch up that makes it allowed in all cases, but only working in a subset of cases | 23:01 |
| zaneb | lifeless: currently we just disallow it because we know it doesn't work | 23:01 |
| SpamapS | zaneb: at what point does this break down? | 23:02 |
| * SpamapS reads the blueprint again :) | 23:02 | |
| zaneb | SpamapS: update after update-failed is broken | 23:02 |
| zaneb | but update after create-failed can work, so we should allow that | 23:02 |
| zaneb | (that's your patch) | 23:03 |
| zaneb | and delete after delete-failed works now, I believe | 23:03 |
| SpamapS | zaneb: ok, so the problem is that if the update fails half-way and we don't want rollback.. we have a template which has been actualized but is not associated with the stack, right? | 23:03 |
| zaneb | update after delete-failed should never work IMO. user lost rights to try to save the stack when they tried to delete it ;) | 23:04 |
| SpamapS | zaneb: so the next update would come along, and try to do the same deltas which have already been done? | 23:04 |
| *** andersonvom has quit IRC | 23:04 | |
| *** dims has quit IRC | 23:04 | |
| SpamapS | zaneb: right, delete failed should just be delete again. | 23:04 |
| zaneb | SpamapS: yeah, so either rollback is disabled or rollback failed. then our internal state is inconsistent | 23:04 |
| SpamapS | zaneb: so what if on update failed and rollback disabled, we just use what is in the database to construct a template to diff against? | 23:05 |
| SpamapS | Or is there not enough in resources to do that? | 23:06 |
| stevebaker | SpamapS, zaneb, I've added a comment to the review | 23:06 |
| zaneb | SpamapS: the resources don't store their template snippets, so we have no way of reconstructing it | 23:06 |
| *** alexpilotti has joined #heat | 23:08 | |
| SpamapS | that would seem a relatively straight forward to handle the situation then. Store the snippets and then when an update starts, assemble the resources section from the database rather than the raw template. | 23:08 |
| SpamapS | that would have a reasonable chance at making update idempotent | 23:09 |
| zaneb | my plan was actually to update the main template one resource at a time | 23:10 |
| SpamapS | same difference really. :) | 23:10 |
| SpamapS | except if you make it per resource you can do it all in parallel. :) | 23:11 |
| *** saurabhs has quit IRC | 23:12 | |
| SpamapS | zaneb: this is nearly "the next thing" for me to tackle.. we're getting the the point where we'll have a long lived stack for our CD cloud.. and we will need an answer for when keystone hiccups during an update :) | 23:12 |
| *** arbylee1 has joined #heat | 23:14 | |
| *** arbylee has quit IRC | 23:15 | |
| *** randallburt has quit IRC | 23:18 | |
| *** dims has joined #heat | 23:20 | |
| *** jcru has quit IRC | 23:20 | |
| *** yogesh has quit IRC | 23:20 | |
| SpamapS | zaneb: well anyway, I completely see your point and I think I'll take your advice, focus my current approach on create failed.. and then expand on it with snippet storing and recovering from an update failed. | 23:21 |
| SpamapS | or I'll test rollback :) | 23:21 |
| *** sballe has quit IRC | 23:22 | |
| *** sballe has joined #heat | 23:22 | |
| *** jergerber has quit IRC | 23:24 | |
| *** randallburt has joined #heat | 23:27 | |
| *** saurabhs has joined #heat | 23:30 | |
| tango | SpamapS: Hi Clint, we have been trying out your patch, we can give a hand in this | 23:33 |
| *** sballe_ has joined #heat | 23:34 | |
| *** sballe has quit IRC | 23:35 | |
| *** jmckind has quit IRC | 23:36 | |
| *** bvandenh has quit IRC | 23:37 | |
| SpamapS | tango: sweet! | 23:39 |
| *** pablosan has quit IRC | 23:45 | |
| *** pablosan has joined #heat | 23:49 | |
| *** andrew_plunk has quit IRC | 23:50 | |
| *** sjmc7 has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!